5-HT1D and 5-HT1E/1F Binding Sites in Depressed Suicides

ORIGINAL RESEARCH ARTICLE

5-HT1D and 5-HT1E/1F binding sites in depressed suicides: increased 5-HT1D binding in globus pallidus but not cortex S Lowther1, CLE Katona2, MR Crompton3 and RW Horton1

1Department of Pharmacology and Clinical Pharmacology; 3Department of Forensic Medicine, St George’s Hospital Medical School, London SW17 0RE; 2Department of Psychiatry, University College London Medical School, London W1N 8AA, UK

5-HT1D and 5-HT1E/1F receptor binding sites were measured in brain samples obtained at post- mortem from suicide victims with a ﬁrm retrospective diagnosis of depression, and matched

controls. In antidepressant-free suicides a signiﬁcantly higher number of 5-HT1D receptors was found in globus pallidus. This was largely restricted to those suicides who died by violent means. This effect was not observed in antidepressant-treated suicides. No differences or

trends in 5-HT1D binding were found in putamen, parietal or frontal cortex, in antidepressant- free or antidepressant-treated suicides. There were no differences in the number of 5-HT1E/1F receptors in any of the regions studied.

Keywords: 5-HT1D receptor binding sites; 5-HT1E/1F receptor binding sites; suicide; depression; post- mortem brain tissue; antidepressant drugs; cerebral cortex; putamen; globus pallidus; basal ganglia

Introduction receptors in a subgroup of subjects with no evidence of depressive symptoms, but not in a smaller subgroup There is abundant evidence that serotonergic function with well documented evidence of depression. is reduced in depression.1,2 Furthermore a consistent A number of 5-HT1D ligands and radioligands have feature of effective treatments is their ability to 10–13 been developed. The 5-HT ␣ and 5-HT ␤ recep- enhance serotonergic transmission.3 1D 1D tors have a very similar pharmacology, but can be dis- Seven classes of 5-HT (5-hydroxytryptamine, tinguished using ketanserin and ritanserin.14 [125I]GTI, serotonin) receptors are recognised; 5-HT have been 1–4 [3H]L-694,247 and [3H]GR 125743 are selective for 5- pharmacologically characterised, 5-HT were ident- 5–7 HT receptors,10,11,15–17 but do not distinguish the two ified by molecular biology and are less well character- 1D forms of the receptor. The 5-HT ␤ receptor appears to ised.4 The 5-HT family is the largest with five subtypes 1D 1 be the dominant 5-HT receptor in human brain. termed 5-HT , 5-HT , 5-HT , 5-HT and 5-HT (the 1D 1A 1B 1D 1E 1F There are no selective ligands for 5-HT and 5-HT 5-HT receptor is now accepted as belonging to the 5- 1E 1F 1C receptors. They are labelled by [3H]5-HT and are HT family, and is termed 5-HT ).4 Two types of 5- 2 2C characterised by their low affinity for 5-carboxamido- HT1D receptor have been reported, termed 5-HT1D␣ and 5,6 tryptamine (5-CT), compared to other 5-HT1 recep- 5-HT ␤. The 5-HT receptor in the rat and mouse 1D 1B tors.18,19 In human frontal cortex these receptors is the species homologue of the 5-HT ␤ receptor in 1D account for over 40% of the total 5-HT receptors.20 humans. 1 They have not been studied to date in relation to Studies of the possible involvement of 5-HT recep- 1 depression. tors in depression have been largely limited to the 5- In the present study we have used [3H]5-HT with HT receptor. The 5-HT receptor is of potential 1A 1D selective pharmacological blockade to examine 5-HT interest because it functions as a nerve terminal autore- 1D and 5-HT receptors (see Methods) in a group of ceptor, regulating the release of 5-HT.7,8 To date only 1E/1F suicides, restricted to those with a firm retrospective one study has examined 5-HT receptors in relation 1D diagnosis of depression. We have studied separately to depression. Arranz et al9 measured 5-HT receptor 1D those subjects who were free of antidepressants for at binding in frontal cortex of suicides. Although they least 3 months prior to death, and those in whom the found no significant difference overall compared to prescription of antidepressant drugs was clearly docu- matched controls, they found lower numbers of 5-HT 1D mented. Four brain regions were studied — frontal and parietal cortex, putamen and globus pallidus. We chose frontal cortex as this region is implicated in depression Correspondence: S Lowther, Department of Pharmacology and 21 Clinical Pharmacology, St George’s Hospital Medical School, and receptor changes have been reported. Another London SW17 0RE, UK. E-mail: slowtherȰsghms.ac.uk cortical region, the parietal cortex, which has not been Received 10 October 1996; revised and accepted 23 January 1997 particularly implicated in depression, was chosen for 5-HT1D binding in depressed suicides S Lowther et al 315 comparative purposes. We have selected two areas of oughly mixed and stored in air-tight containers at the basal ganglia for study since structural and meta- −80°C until assayed. Brain areas studied were frontal bolic changes have been reported in affective dis- cortex (Brodmann area 10), parietal cortex (Brodmann orders.22–24 In human brain, the highest density of 5- area 7), putamen and globus pallidus. 20 HT1D receptors is found in globus pallidus, making it potentially interesting in relation to depression. Tissue preparation Membranes were prepared essentially as described by Cheetham et al.26 Frozen brain tissue was homogenised Materials and methods in ice-cold 0.25 M sucrose (1 : 20 w/v) using a motor- Subject selection driven teflon pestle (eight strokes at 120 rpm), and cen- Deaths recorded as suicide at Coroners’ inquest were trifuged (at 4°C) at 1000 g for 10 min. The supernatant subjected to retrospective diagnosis by a psychiatrist was stored on ice and the pellet rehomogenised in (CLEK), using hospital and Coroners’ records and inter- 0.25 M sucrose (1 : 10 w/v) and centrifuged at 750 g for views with the subject’s medical practitioner. Suicides 10 min. Combined supernatants were diluted with were divided into five groups according to the classi- 50 mM Tris-HCl buffer, pH 7.4 (at 25°C) to 1 : 100 w/v fication of Beskow et al.25 The subjects included in this and centrifuged at 35 000 g for 10 min. The membrane study came from group 1 (endogenous depression) and pellet was resuspended in 50 mM Tris-HCl buffer group 2 (depressive syndrome) only. Subjects from (1 : 50 w/v) and incubated at 37°C for 10 min to remove groups 3, 4 and 5, with diagnoses of schizophrenia, endogenous 5-HT, then centrifuged at 35 000 g for 10 personality disorder, epilepsy, alcohol or other drug min. The pellet was washed by resuspension in 50 mM abuse, alone or in combination with depression, and Tris-HCl buffer (1 : 100 w/v) and centrifuged at those with no clear psychiatric diagnosis, were 35 000 g for 10 min. The final pellet was resuspended excluded. Information on current and previous drug in 50 mM Tris-HCl buffer at a concentration of 5.3 mg treatment was sought and blood samples obtained at original wet weight ml−1 for cortical areas and post-mortem were analysed for the presence of psy- 2.6 mg ml−1 for putamen and globus pallidus, and choactive drugs. This information allowed two distinct immediately used in the binding assay. groups of suicides to be studied. Group 1. Antidepressant-free suicides (n = 20). These Binding assays

subjects had not been prescribed antidepressant drugs Total 5-HT1, 5-HT1D/1E/1F and 5-HT1E/1F binding was within 3 months prior to death. Causes of death were determined concurrently (as previously described).20 hanging (n = 8), carbon monoxide poisoning (n = 5), Assays were performed in 5-ml plastic tubes. Drug sol- drug overdosage (n = 5), self-inflicted wounding (n = 1) utions were prepared in water, and ligand in 50 mM and jumping from height (n = 1). Control brains for this Tris-HCl buffer, pH 7.4 (at 25°C). Common to all assays group were from subjects who died suddenly from were 500 ␮l freshly prepared membranes, 100 ␮lof causes not involving the central nervous system, and 50 mM Tris-HCl buffer, pH 7.4 containing 4 mM cal- without documented evidence of mental illness. Con- cium chloride (final concentration) and 0.1% ascorb- trols were individually matched for age (75% within ate, 100 ␮l water (total binding) or unlabelled 5-HT 5 years) and gender. Causes of death were myocardial (10 ␮M, to define non-specific binding) and 100 ␮l infarction (n = 17), accidental fall (n = 1), acute asthma [3H]5-HT (25.4 Ci mmol−1, NEN Dupont, Boston, MA, (n = 1) and road traffic accident (n = 1). USA) at eight concentrations (0.5–32 nM final). 5- = Group 2. Antidepressant-treated suicides (n 16). HT1D/1E/1F binding sites were characterised by the pres- Prescription of antidepressant drugs was clearly docu- ence of 100 ␮l(−)pindolol (500 nM final) and 100 ␮l

mented in this group. Details of drug treatment and mesulergine (100 nM final) to block 5-HT1A and 5-HT2C causes of death for these suicides are shown in Table 1. receptors respectively. (5-HT1B receptors are absent in 27 A second control group was individually matched for human brain.) 5-HT1E/1F binding was characterised by age (100% within 5 years) and gender to the antide- 100 ␮l 5-carboxamidotryptamine (5-CT, 100 nM final) ␮ pressant-treated suicides. Causes of death for controls and 100 l mesulergine (100 nM final) to block 5-HT1A, = were myocardial infarction (n 12), road traffic acci- 5-HT1D and 5-HT2C receptors. 5-HT1D binding was = = dent (n 2), drowning (n 1) and viral pneumonia determined by subtraction of the Bmax for 5-HT1E/1F = (n 1). from 5-HT1D/1E/1F binding. Two hundred microlitres of water replaced pindolol/5-CT and mesulergine to give

Tissue collection, dissection and storage total 5-HT1 binding. Tissue collection, dissection and storage were as pre- Samples were mixed thoroughly and incubated at viously described by Cheetham et al.26 Briefly, brains 37°C for 20 min. Membrane-bound radioactivity was were obtained within 78 h of death and stored at −80°C. recovered by filtration under vacuum through What- Eighteen hours prior to dissection, brains were trans- man GF/C glass fiber filters using a Brandel cell har- ferred to −20°C. Coronal sections (3 mm thick) were cut vester (Gaithersburg, USA). Filters were washed with and dissected on a perspex surface cooled with carbon 16 ml ice-cold 50 mM Tris-HCl buffer. Radioactivity dioxide granules. Great care was taken to include only was determined by liquid scintillation counting using grey matter in cortical samples. Dissected areas were Packard scintillator 299 at an efficiency of 39–44%. finely chopped into approximately 3-mm cubes, thor- Aliquots of membrane were stored at −20°C for sub- 5-HT1D binding in depressed suicides S Lowther et al 316 Table 1 Drug treatment and cause of death for the antidepressant drug-treated suicides

Drug treatment Duration of treatment Cause of death (weeks)

Overdosage with antidepressant drugs Trimipramine, Temazepam, Diazepam 2 Trimipramine, Aspirin Dothiepin, Haloperidol 4 Dothiepin, Haloperidol Amitriptyline, Diazepam 20 Amitriptyline, Aspirin Dothiepin 38 Dothiepin Imipramine, Lithium, Diazepam 40 Imipramine Amitriptyline Ͼ52 Amitriptyline Dothiepin Ͼ52 Dothiepin, Paracetamol Trimipramine, Nitrazepam Ͼ52 Trimipramine, Aspirin, Nitrazepam

Overdosage without antidepressant drugs Mianserin 4 Barbiturate, Temazepam Clomipramine, Thioridazine, Procyclidine 16 Dextropropoxyphene, Paracetamol Imipramine, Flupenthixol Ͼ52 Aspirin

Violent methods Amitriptyline 1 Hanging Amitriptyline 1 Hanging Dothiepin, Temazepam 3 Jumping from height Clomipramine, Lithium 4 Hanging Amitriptyline, Triﬂuoperazine, Temazepam 11 Hanging

sequent protein determination,28 using bovine serum post-mortem delay (time from death to storage of tissue albumin as the standard. Assays were performed on at −80°C) or tissue storage time (time from storage at coded samples, blind to subject classiﬁcation, but −80°C to assay) between the suicide groups and their arranged so that suicide and controls were assayed respective matched controls. concurrently. Tissue availability was such that it was

not possible to study all brain regions in all subjects. 5-HT1D and 5-HT1E/1F binding in antidepressant-free suicides and controls Analysis There were no signiﬁcant correlations between age or

The equilibrium dissociation constant (Kd) and post-mortem delay and binding parameters. 3 maximum number of binding sites (Bmax) were deter- Bmax values for [ H]5-HT binding are shown in mined by non-linear regression ﬁtting to a one-site Table 3. The number of total 5-HT1, 5-HT1D and 5- binding model. Comparisons between groups were HT1E/1F binding sites did not differ between suicides made using unpaired t-test of log transformed data with and controls in putamen, frontal cortex and parietal Bonferroni’s multiple comparison test. Correlations cortex. However in globus pallidus the number of 5-

were determined using Pearson’s correlation. HT1D sites was signiﬁcantly higher in suicides than = Ͻ controls (by 24%, t 2.52, P 0.05), whereas 5-HT1E/1F sites did not differ (t = 0.13, ns, Table 3). Individual Results Bmax values in globus pallidus are shown in Figure 1. Demographic details of subjects are shown in Table 2. When suicides were divided according to violence of

There were no signiﬁcant differences in age, gender, death, the higher density of 5-HT1D sites seen in globus

Table 2 Demographic details of subjects studied

n Age (years) Gender Post-mortem delay Storage time (h) (months)

Controls 20 43 ± 3 (20–60) 15 M, 5 F 40 ± 3 (16–72) 65 ± 3 (24–78) Antidepressant-free suicides 20 43 ± 3 (20–60) 15 M, 5 F 31 ± 3 (15–65) 61 ± 4 (36–81)

Controls 16 44 ± 4 (18–69) 8 M, 8 F 44 ± 5 (23–78) 42 ± 6 (12–78) Antidepressant-treated suicides 16 44 ± 4 (16–68) 8 M, 8 F 39 ± 4 (5–73) 49 ± 5 (13–87)

Data expressed as means ± s.e.m., range shown in parentheses. Tissue storage time relates to frontal cortical samples, storage time for other brain regions differed by no more than ±4 months. None of the comparisons between suicides and controls reached statistical signiﬁcance (P Ͻ 0.05, unpaired t-test). 5-HT1D binding in depressed suicides S Lowther et al 317 3 Table 3 Bmax values for [ H]5-HT binding in antidepressant-free suicides and controls

Brain region n Total 5-HT1 sites 5-HT1D/1E/1F 5-HT1E/1F 5-HTID (by subtraction)

Putamen Control 19 253 ± 8 193 ± 10 133 ± 860±6 Suicide 19 250 ± 13 204 ± 12 143 ± 10 61 ± 5 Globus pallidus Control 18 353 ± 22 328 ± 23 107 ± 9 221 ± 17 Suicide 20 423 ± 19 377 ± 17 104 ± 7 273 ± 17* Frontal cortex Control 18 308 ± 11 193 ± 6 106 ± 588±5 Suicide 17 300 ± 10 196 ± 8 108 ± 689±7 Parietal cortex Control 16 206 ± 11 134 ± 779±655±7 Suicide 20 215 ± 9 145 ± 891±554±6

−1 ± = Bmax, fmol mg protein. Values are means s.e.m. n number of subjects. Signiﬁcant differences between controls and suicides are denoted by *P Ͻ 0.05 (unpaired t-test with Bonferroni’s multiple comparison test).

No signiﬁcant differences were found in Kd values between antidepressant-free suicides and matched controls in any of the brain regions studied (Table 4).

5-HT1D and 5-HT1E/1F binding in antidepressant- treated suicides and controls Signiﬁcant negative correlations were found between

post-mortem delay and total 5-HT1 Bmax (putamen, = = P 0.05) and 5-HT1D Bmax (globus pallidus, P 0.009). In parietal cortex, signiﬁcant negative correlations

were found between age and total 5-HT1 Bmax = = (P 0.016) and 5-HT1E/1F Bmax (P 0.035). There were no signiﬁcant correlations between age or post-mortem

delay and Kd values. 3 Bmax values for [ H]5-HT binding are shown in Table 5. Bmax values did not differ between antidepressant-treated suicides and controls in any of the regions studied, or when suicides were divided according to violence of death. There was considerable vari- ation in the duration of antidepressant treatment (Table 1). Some subjects had relatively short treatment (up to 4 weeks, n = 7) and some had received longer treatment (over 11 weeks, n = 9). Even when only those subjects with longer antidepressant treatment were

considered, there were still no differences in Bmax values compared to matched controls.

Kd values for antidepressant-treated suicides showed more variability than for antidepressant-free suicides

Figure 1 Individual Bmax values for 5-HT1D and 5-HT1E/1F (Table 4). Decreased binding affinities (increased Kd binding sites in globus pallidus of antidepressant-free sui- values) for total 5-HT1 binding were seen in frontal cor- cides and controls. Significant differences between controls = Ͻ Ͻ tex (30%, t 2.74, P 0.01) and parietal cortex (43%, and suicides are denoted by *P 0.05 (unpaired t-test with t = 2.68, P Ͻ 0.05). These increased K values were seen Bonferroni’s multiple comparison test). d in those suicides dying by drug overdosage (ie non- violently) and were absent in those dying by violent pallidus (Figure 2) was restricted to suicides who died means. by violent means (by 38%, t = 2.69, P Ͻ 0.05, suicides = = n 10, controls n 9). Non-violent suicides did not dif- Discussion fer from their matched controls (t = 0.97, ns, suicides = = n 10, controls n 9). Bmax values for 5-HT1D and 5- In the present study we have attempted to examine the HT1E/1F sites did not differ between violent or non- potential role of some of the more recently defined 5- violent suicides and their matched controls in HT1 receptor subtypes in depression. Although there putamen, frontal cortex or parietal cortex. are relatively selective radioligands for 5-HT1D recep- 5-HT1D binding in depressed suicides S Lowther et al 318

Figure 2 5-HT1D binding sites in globus pallidus and frontal cortex of violent and non-violent antidepressant-free suicides. Values are means ± s.e.m. Signiﬁcant differences between controls and suicides are denoted by *P Ͻ 0.05 (unpaired t-test with Bonferroni’s multiple comparison test).

3 Table 4 Kd values for [ H]5-HT binding in antidepressant-free and antidepressant-treated suicides and controls

Brain region Antidepressant-free suicides Antidepressant-treated suicides

Total 5-HT1 5-HT1D/1E/1F 5-HT1E/1F Total 5-HT1 5-HT1D/1E/1F 5-HT1E/1F

Putamen Control 3.32 ± 0.19 4.03 ± 0.31 4.23 ± 0.31 3.21 ± 0.19 3.63 ± 0.28 3.24 ± 0.42 Suicide 3.61 ± 0.29 4.27 ± 0.22 4.33 ± 0.27 4.26 ± 0.38 4.69 ± 0.27 4.19 ± 0.39 Globus pallidus Control 2.96 ± 0.21 4.18 ± 0.30 6.61 ± 0.55 2.24 ± 0.19 3.08 ± 0.32 5.61 ± 0.48 Suicide 2.97 ± 0.15 4.01 ± 0.20 6.04 ± 0.43 2.65 ± 0.30 3.14 ± 0.19 3.80 ± 0.62 Frontal cortex Control 2.48 ± 0.16 4.44 ± 0.25 3.39 ± 0.25 2.13 ± 0.10 3.83 ± 0.29 4.18 ± 0.36 Suicide 2.60 ± 0.23 4.59 ± 0.23 3.52 ± 0.22 2.76 ± 0.17* 4.43 ± 0.26 3.95 ± 0.32 Parietal cortex Control 2.96 ± 0.22 4.30 ± 0.23 3.51 ± 0.34 2.61 ± 0.14 4.23 ± 0.26 4.79 ± 0.43 Suicide 3.33 ± 0.29 5.13 ± 0.32 3.82 ± 0.34 3.73 ± 0.40 4.78 ± 0.35 5.16 ± 0.29

± = Kd nM. Values are means s.e.m. n 16–20 for antidepressant-free suicides and controls, and 14–16 for antidepressant-treated suicides and controls. Signiﬁcant differences between controls and suicides are denoted by *P Ͻ 0.05, (unpaired t-test with Bonferroni’s multiple comparison test).

125 3 3 tors ([ I]GTI, [ H]L-694,247 and [ H]GR receptors. We have not attempted to measure 5-HT1E 10,11,15–17 125743), there are no selective radioligands for and 5-HT1F receptors separately, although this would 5-HT1E and 5-HT1F receptors. 5-HT1E/1F receptors can be possible by using sumatriptan. Although we have 3 be measured together by labelling with [ H]5-HT and quantiﬁed 5-HT1D receptor number by subtraction utilising their low afﬁnity for 5-CT, which distingu- using [3H]5-HT rather than a selective radioligand, the

ishes them from other 5-HT1-like sites. 5-HT1F recep- values we obtained were robust and reproducible. tors can be distinguished from 5-HT1E receptors by We have measured 5-HT1D and combined 5-HT1E/1F their high affinity for sumatriptan, and thus 5-HT1F receptors in brain samples obtained at post-mortem receptors can be radiolabelled with [3H]sumatriptan in from suicides with a firm retrospective diagnosis of the presence of low concentrations of 5-CT to occupy depression. We have demonstrated a significantly 19,29 5-HT1D receptors. Our approach, based on Lowther higher number of 5-HT1D receptors in the globus pal- et al20 was to use [3H]5-HT in the presence of lidus of antidepressant-free suicides. There were no − ( )pindolol and mesulergine (to block 5-HT1A/2C differences, or even trends suggestive of such differ- receptors) to measure combined 5-HT1D/1E/1F receptors. ences, in the numbers of 5-HT1E/1F receptors in globus 5-CT was used to distinguish 5-HT1D from 5-HT1E/1F pallidus, or of any receptors measured in other regions. 5-HT1D binding in depressed suicides S Lowther et al 319 3 Table 5 Bmax values for [ H]5-HT binding in antidepressant-treated suicides and controls

Brain region n Total 5-HT1 sites 5-HT1D/1E/1F 5-HT1E/1F 5-HT1D (by subtraction)

Putamen Control 14 203 ± 9 161 ± 9 105 ± 755±7 Suicide 16 215 ± 12 175 ± 12 108 ± 767±8 Globus pallidus Control 15 386 ± 25 340 ± 25 100 ± 7 240 ± 21 Suicide 14 400 ± 18 345 ± 20 99 ± 11 246 ± 14 Frontal cortex Control 14 250 ± 7 164 ± 7 102 ± 562±8 Suicide 15 256 ± 13 171 ± 996±675±9 Parietal cortex Control 16 211 ± 8 147 ± 693±553±5 Suicide 16 224 ± 12 150 ± 794±656±6

−1 ± = Bmax, fmol mg protein. Values are means s.e.m. n number of subjects. None of the comparisons between controls and suicides reached statistical signiﬁcance (P Ͻ 0.05, unpaired t-test with Bonferroni’s multiple comparison test).

We need to consider the factors that might influence thus be a reflection of the small proportion of suicides this apparently selective increase in 5-HT1D receptors. within this group who died by violent means. Suicides and controls were well matched for age and Our approach does not allow the determination of 3 post-mortem delay, and we found no evidence for a the Kd of [ H]5-HT binding to 5-HT1D receptors. In anti- relationship between 5-HT1D receptors and these vari- depressant-free suicides, there were no differences in ables in antidepressant-free suicides, either in globus Kd values for total 5-HT1, 5-HT1D/1E/1F and 5-HT1E/1F pallidus or the other brain areas studied. The higher receptors in any region. In contrast, Kd values were number of 5-HT1D receptors was almost entirely restric- generally higher, although not markedly or consist- ted to those suicides who died by violent means ently so, in antidepressant-treated suicides (Table 4). (Figure 2). This discounts the possibility of an effect These findings are consistent with an interaction of related to the ingestion of toxic agents. We therefore antidepressant drugs (either in treatment or in feel confident that this difference has not arisen arte- overdosage) with 5-HT1 receptors. factually. The only previous study to examine 5-HT1D recep- Within those subjects who had been prescribed anti- tors in suicide was restricted to the frontal cortex.9 The depressants, there were no differences in the number number of 5-HT1D receptors did not differ in a sub- of 5-HT1D and 5-HT1E/1F receptors in any of the brain group of suicides with documented evidence of regions studied. The lack of difference in 5-HT1D bind- depression, or in suicides who died violently. These ing in the globus pallidus in this group contrasts with results are consistent with the present study. Arranz 9 the increased 5-HT1D binding found in the antidepress- et al did however report lower 5-HT1D binding in a ant-free suicides. A possible explanation is that antide- subgroup of suicides who had no evidence of depress- pressant treatment may have reduced the number of 5- ive symptoms (from medical records and information

HT1D receptors. This is in keeping with evidence in the from closest relatives). The present study was restric- guinea pig that the nerve terminal 5-HT autoreceptor ted to suicides with a ﬁrm retrospective diagnosis of

(5-HT1D) densensitises in hypothalamus and hippo- depression, and therefore we are unable to comment campus during long-term administration of some anti- further on the findings of Arranz et al.9 depressant drugs.30 However no studies appear to have Despite the growing number of morphological and in examined the effect of antidepressants on 5-HT1D bind- vivo imaging studies implicating the basal ganglia in ing. Also if antidepressants decrease 5-HT1D binding, affective disorders, few neurochemical studies have one would have expected to see reduced binding in examined these regions. We have recently studied other brain regions; this clearly did not occur. On bal- dopamine D1 and D2 receptors and found them to be ance it is therefore unlikely that our results can be attri- unaltered in caudate, putamen and nucleus accumbens buted to the effects of antidepressant treatment. of depressed suicides.31 We have also found unaltered A more plausible explanation relates to the pro- 5-HT uptake sites, labelled with [3H]paroxetine, in cau- portion of suicides dying by violent and non-violent date and putamen of similar subjects.32 means. Within the antidepressant-treated suicides, The increase in 5-HT1D receptors in globus pallidus results in globus pallidus were obtained from five sui- was rather a surprise finding and one that cannot easily cides who died violently. These subjects showed 20% be incorporated into current neuroanatomical sub- higher 5-HT1D binding in the suicides (controls strates of depression. It is known from animal studies ± ± −1 216 31, suicides 259 7 fmol mg protein), whereas that 5-HT1D receptors act as nerve terminal autorecep- 7,8 suicides dying non-violently did not differ from con- tors for 5-HT. The increase in 5-HT1D receptors might trols. The lack of difference in 5-HT1D binding in glo- imply an increased serotonergic innervation of the globus pallidus in antidepressant-treated suicides may bus pallidus in depression. However we believe this 5-HT1D binding in depressed suicides S Lowther et al 320 not to be the case, because we found a trend for 5-HT 5 Hartig PR, Branchek TA, Weinshank RL. A subfamily of 5-HT1D uptake sites (a presynaptic marker of 5-HT neurones) receptor genes. Trends Pharmacol Sci 1992; 13: 152–159. 6 Weinshank RL, Zgombick JM, Macchi M, Branchek TA, Hartig PR. to be lower in globus pallidus of antidepressant-free Human serotonin 1D receptor is encoded by a subfamily of two ± −1 depressed suicides (controls 230 19 fmol mg pro- distinct genes: 5-HT1D␣ and 5-HT1D␤. Proc Natl Acad Sci USA 1992; − tein, n = 20; suicides 193 ± 11 fmol mg 1 protein, 89: 3630–3634. n = 20).33 If 5-HT uptake sites indicate lower 5-HT 7 Hoyer D, Middlemiss DN. Species differences in the pharmacology innervation then it could be argued that the increase of terminal 5-HT autoreceptors in mammalian brain. Trends Phar- macol Sci 1989; 10: 131–132. in 5-HT1D receptors is an adaptive post-synaptic 8 Maura G, Thellung ST, Andriolo GC, Ruelle A, Raiteri M. Release-

response to reduced 5-HT release. regulating serotonin 5-HT1D receptors in human cerebral cortex. J No previous studies have examined 5-HT1E/1F recep- Neurochem 1993; 60: 1179–1182. tors in depression. Although the function of these 9 Arranz B, Eriksson A, Mellerup E, Plenge P, Marcusson J. Brain 5- HT , 5-HT , and 5-HT receptors in suicide victims. Biol Psy- receptors has yet to be elucidated, our findings do not 1A 1D 2 chiatry 1994; 35: 457–463. indicate a critical involvement in depression and/or 10 Boulenguez P, Chauveau J, Segu L, Morel A, Delaage M, Lanoir suicide. J. Pharmacological characterization of serotonin-O-carboxymethyl- Other 5-HT receptors have been more extensively glycyl-tyrosinamide, a new selective indolic ligand for 5-hydroxytryptamine (5-HT) and 5-HT binding sites. J Pharmacol Exp studied in relation to depression, particularly 5-HT1A 1B 1D and 5-HT receptors. No differences in 5-HT recep- Ther 1991; 259: 1360–1365. 2A 1A 11 Bruinvels AT, Landwehrmeyer B, Waeber C, Palacios JM, Hoyer D. tors have been reported in suicides (or violent suicides) Homogeneous 5-HT recognition sites in the human substantia 9,34,35 1D and controls in homogenate binding studies, nigra identified with a new iodinated radioligand. Eur J Pharmacol although two studies using autoradiography have 1991; 202: 89–91. shown localised increases in hippocampus and cor- 12 Beer MS, Stanton JA, Hawkins LM, Middlemiss DN. 5-Carboxami- 36,37 dotryptamine-insensitive 5-HT1-like receptors are concentrated in tex. Despite greater emphasis on 5-HT2A receptors guinea pig but not rat, claustrum. Eur J Pharmacol 1993; 236: in depression, the findings remain controversial with 167–169. some studies reporting increased receptor numbers in 13 Skingle M, Skopes DIC, Feniuk W, Connor HE, Carter MC, Cli-

the cortex of suicides, particularly in those dying by therow MC. GR 127935: a potent orally active 5-HT1D receptor violent means, compared to controls,38–42 while others antagonist. Br J Pharmacol 1993; 110: 9P. have not found significant differences.26,43,44 To the 14 Zgombick JM, Schechter LE, Kucharewicz SA, Weinshank RL, best of our knowledge, no studies have examined 5- Branchek TA. Ketanserin and ritanserin discriminate between recombinant human 5-HT1D␣ and 5-HT1D␤ receptor subtypes. Eur J HT1A and 5-HT2A receptors in globus pallidus making Pharmacol (Mol Pharmacol) 1995; 291: 9–15. it difficult to judge the relevance of the present find- 15 Beer MS, Middlemiss DN. Serotonin-5-O-carboxymethyl-glycyl[125-

ings. I]tyrosinamide labels the 5-HT1D␤ receptor subtype in human cortex. Eur J Pharmacol 1993; 242: 195–198. In summary we have found increased 5-HT1D receptor binding in globus pallidus of antidepressant-free 16 Bruinvels AT, Landwehrmeyer B, Probst A, Palacios JM, Hoyer D. A comparative autoradiographic study of 5-HT1D binding sites in suicides. This higher number of 5-HT1D receptors was human and guinea-pig brain using different radioligands. Mol Brain not apparent in antidepressant-treated suicides, but we Res 1994; 21: 19–29. have argued that this is a reflection of the smaller pro- 17 Heald A, Stanton JA, Osborne S-A, Middlemiss DN, Beer MS. 3 portion of subjects dying by violent means, rather than [ H]L-694,247 labels the 5-HT1D␤ receptor in pig caudate mem- the effects of antidepressant treatment. The relation- branes. Eur J Pharmacol 1994; 264: 213–216. 18 Leonhardt S, Herrick-Davis K, Titeler M. Detector of a novel sero- ship between the functional consequences of increased tonin receptor subtype (5-HT1E) in human brain: interaction with 5-HT1D binding within the globus pallidus and a GTP-binding protein. J Neurochem 1989; 53: 465–471. depression or specific depressive symptomatology is 19 Adham N, Kao H-T, Schechter LE, Bard J, Olsen M, Urquhart D, not clear. Durkin M, Hartig PR, Weinshank RL, Branchek TA. Cloning of another human serotonin receptor (5-HT1F): a fifth 5-HT1 receptor subtype coupled to the inhibition of adenylate cyclase. Proc Natl Acknowledgements Acad Sci USA 1993; 90: 408–412. 20 Lowther S, De Paermentier F, Crompton MR, Horton RW. The dis- We gratefully acknowledge a research award from the tribution of 5-HT1D and 5-HT1E binding sites in human brain. Eur Theodore and Vada Stanley Foundation. J Pharmacol 1992; 222: 137–142. 21 Horton RW. The neurochemistry of depression: evidence derived from studies of post-mortem brain tissue. Mol Aspects Med 1992; 13: 191–203. References 22 Krishnan KRR, McDonald WM, Escalona PR, Doraiswamy PM, Na 1 Owens MJ, Nemeroff CB. Role of serotonin in the pathophysiology C, Husain MM, Figiel GS, Boyko OB, Ellinwood EH, Nemeroff CB. of depression: focus on the serotonin transporter. Clin Chem 1994; Magnetic resonance imaging of the caudate nuclei in depression. 40: 288–295. Arch Gen Psychiatry 1992; 49: 553–557. 2 Deakin JFW. 5-HT, antidepressant drugs and the psychosocial ori- 23 Aylward EH, Roberts-Twillie JV, Barta PE, Kumar AJ, Harris GJ, gins of depression. J Psychopharmacol 1996; 10: 31–38. Geer M, Peyser CE, Pearlson GD. Basal ganglia volumes and white 3 De Montigny C, Blier P. Effects of antidepressant treatment on 5- matter hyperintensities in patients with bipolar disorder. Am J Psy- HT neurotransmission. In: Marsden CA, Heal DJ (eds). Central chiatry 1994; 151: 687–693. Serotonin Receptors and Psychotropic Drugs. Blackwell Scientific 24 Moore CM, Christensen JD, Lafer B, Fava M, Renshaw PF. Lower Publications: Oxford, 1992, pp 223–238. levels of nucleoside triphosphate in the basal ganglia of depressed 4 Hoyer D, Clarke DE, Fozard JR, Hartig PR, Martin GR, Mylecharane subjects: a phosphorus-31 magnetic resonance spectroscopy study. EJ, Saxena PR, Humphrey PP. International Union of Pharmacology Am J Psychiatry 1997; 154: 116–118. classification of receptors for 5-hydroxytryptamine (serotonin). 25 Beskow J, Gottfries CG, Roos BE, Winblad B. Determination of Pharmacol Rev 1994; 46: 157–203. monoamine and monoamine metabolites in the human brain: post 5-HT1D binding in depressed suicides S Lowther et al 321 mortem studies in a group of suicides and in a control group. Acta CLE, Horton RW. 5-HT1A receptor binding sites in post-mortem Psychiatr Scand 1976; 53: 7–20. brain samples from depressed suicides and controls. J Affect Disord

26 Cheetham SC, Crompton MR, Katona CLE, Horton RW. Brain 5-HT2 1997; (in press). receptor binding sites in depressed suicide victims. Brain Res 1988; 36 Joyce JN, Shane A, Lexow N, Winokur A, Casanova MF, Kleinman 443: 272–280. JE. Serotonin uptake sites and serotonin receptors are altered in the 27 Hoyer D, Pazos A, Probst A, Palacios JM. Serotonin receptors in limbic system of schizophrenics. Neuropsychopharmacology 1993; human brain. I. Characterization and autoradiographic localization 8: 315–336.

of 5-HT1A recognition sites. Apparent absence of 5-HT1B recog- 37 Arango V, Underwood MD, Gubbi AV, Mann JJ. Localized alter- nition sites. Brain Res 1986; 376: 85–96. ations in pre- and postsynaptic serotonin binding sites in the ven- 28 Lowry OH, Rosebrough NJ, Farr AL, Randall RJ. Protein measure- trolateral prefrontal cortex of suicide victims. Brain Res 1995; 688: ment with the Folin phenol reagent. J Biol Chem 1951; 193: 265– 121–133. 275. 38 Stanley M, Mann JJ. Increased serotonin-2 binding sites in frontal 3 29 Waeber C, Moskowitz MA. [ H]Sumatriptan labels both 5-HT1D and cortex of suicide victims. Lancet 1983; 1: 214–216. 5-HT1F receptor binding sites in the guinea pig brain: an autoradio- 39 Mann JJ, Stanley M, McBride A, McEwan BS. Increased serotonin2 graphic study. Naunyn-Schmiedeberg’s Arch Pharmacol 1995; 352: and ␤-adrenergic receptor binding in the frontal cortices of suicide 263–275. victims. Arch Gen Psychiatry 1986; 43: 954–959. 30 Blier P, Bouchard C. Modulation of 5-HT release in the guinea-pig 40 Arora RC, Meltzer HY. Serotonergic measures in the brains of suic-

brain following long-term administration of antidepressant drugs. ide victims: 5-HT2 binding sites in the frontal cortex of suicide vic- Br J Pharmacol 1994; 113: 485–495. tims and control subjects. Am J Psychiatry 1989; 146: 730–736. 31 Bowden C, Theodorou AE, Cheetham SC, Lowther S, Katona CLE, 41 Arango V, Ernssberger P, Marzuk PM, Chen J-S, Tierney H, Stanley

Crompton MR, Horton RW. Dopamine D1 and D2 receptor binding M, Reis DJ, Mann JJ. Autoradiographic demonstration of increased ␤ sites in brain samples from depressed suicides and controls. Brain serotonin 5-HT2 and -adrenergic binding sites in the brain of suic- Res 1997; (in press). ide victims. Arch Gen Psychiatry 1990; 47: 1038–1047. 32 Lawrence KM, De Paermentier F, Cheetham SC, Crompton MR, 42 Hrdina PD, Demeter E, Vu TB, So´to´nyi P, Palkovits M. 5-HT uptake

Katona CLE, Horton RW. Brain 5-HT uptake sites, labelled with sites and 5-HT2 receptors in brain of antidepressant-free suicide 3 [ H]paroxetine, in antidepressant-free depressed suicides. Brain victims/depressives: increases in 5-HT2 sites in cortex and amyg- Res 1990; 526: 17–22. dala. Brain Res 1993; 614: 37–44. 33 Lawrence KM. The 5-hydroxytryptamine uptake site in depression: 43 Owen F, Chambers DR, Cooper SJ, Crow TJ, Johnson JA, Lofthouse studies of blood platelets and post-mortem brain tissue. PhD thesis, R, Poulton M. Serotonergic mechanisms in brains of suicide vic- University of London, 1991. tims. Brain Res 1986; 362: 185–188. 34 Matsubara S, Arora RC, Meltzer HY. Serotonergic measures in suic- 44 Lowther S, De Paermentier F, Crompton MR, Katona CLE, Horton

ide brain: 5-HT1A binding sites in frontal cortex of suicide victims. RW. Brain 5-HT2 receptors in suicide victims: violence of death, J Neural Transm [GenSect] 1991; 85: 181–194. depression and effects of antidepressant treatment. Brain Res 1994; 35 Lowther S, De Paermentier F, Cheetham SC, Crompton MR, Katona 642: 281–289.